Jump to content

Why are my results here so much lower than with other tests?


UnfriendlyBG

Recommended Posts

I don't think this is accurate.  I'm looking at the windows resource monitor while i do tests elsewhere (speedtest.net, the one from my isp, shaperprobe) and they will get up to 70 Mbps in the resource monitor.  I'm not even talking about the final number im talking about what windows is saying the traffic coming to/going from my computer is.  I can use shaperprobe and it connects to a server somewhere around Miami which isnt close to me and see it get a constant 70mpbs on the windows resource monitor.

 

The speed test here says im getting 22 Mbps down, which just isnt true, ive downloaded stuff at a higher rate than that. im supposed to be getting 105 down.  Im trying to troubleshoot an issue with my upload.  This site says I have between 2-5mpbs upload.  I should be getting 20. Again I think the results are on the low side. I think there is some sort of issue with my upload but its not 2mbps regularly. I stream content at 2.5mpbs all the time.  I'm not saying that speedtest.net and my isps test are accurate, im just questioning the accuracy of this one when its obviously inaccurate in my case. 

 

I got the same results trying the multithreading test so its not that.  I just dont get why I can run other tools while monitoring the network bandwidth and they go much much higher, and they arent just on my isp's network.

Link to comment
Share on other sites

First, try the multithread speed test ... speedtest.net multithreads, TestMy.net by default is linear. Loading much as a single file transfer would. But I do offer the multithread option. If you get drastically differing results between the two tests it can be an indication that there is an underlying issue. There is a wide range of issues... from improper MTU and RWIN settings to a bad cable modem. Seriously, if you're getting lower results with TestMy.net it can be something as obious as a bad modem... and other speed tests WILL NOT DETECT IT. It DOES affect your performance... if your performance is affected in my tests your performance online is affected. Usually a noticeable difference once you find the issue. (looking at your results, I see you've already used the multithread option and got over 50 Mbps, something is restricting that linear transfer... everyone else that's running correctly is able to get much higher scores, something is limiting you.)

Don't be so quick to write off the test. There are many people who have put my methodology under the microscope and consider this to be the most accurate speed test online. (e.g. "I can say with 100% certainty that you will not find a testing methodology more accurate than testmy.net") You really can't compare TestMy.net results to a test that by their own admission makes adjustments to the results. Reference Why Do My Results Differ From Speedtest.net / Ookla Speed Tests?

Look closer at that bandwidth meter. There are anomalies with my competitors even under the best circumstances. Realize too that most flash speed tests are multithreading... you shouldn't have to multithread to get your full speed. An ideal connection will perform good either way... mine does, look.

Multithread

:::.. Download Speed Test Result Details ..:::

Download Connection Speed:: 108266 Kbps or 108.3 Mbps

Download Speed Test Size:: 200.9 MB or 205692 kB or 210628710 bytes

Download Binary File Transfer Speed:: 13533 kB/s or 13.5 MB/s

Tested At:: http://TestMy.net Version 13

Validation:: https://testmy.net/db/qbQ34Jc

More Stats:: https://testmy.net/quickstats/CA3LE https://testmy.net/compID/69431494760

Test Time:: 2014-02-14 11:21:44 Local Time

Location:: Monument, CO US >> Destination:: Global Multithread

1MB Download in 0.08 Seconds - 1GB Download in ~1 Minutes - 1933X faster than 56K

This test of exactly 205692 kB took 15.564 seconds to complete

Running at 482% of hosts average (Comcast Cable https://testmy.net/hoststats/comcast_cable)

User Agent:: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/32.0.1700.77 Safari/537.36 [!]

Default Linear

:::.. Download Speed Test Result Details ..:::

Download Connection Speed:: 101707 Kbps or 101.7 Mbps

Download Speed Test Size:: 200 MB or 204800 kB or 209715200 bytes

Download Binary File Transfer Speed:: 12713 kB/s or 12.7 MB/s

Tested At:: http://TestMy.net Version 13

Validation:: https://testmy.net/db/6hgXTYD

TiP Measurement Summary:: Min 47.84 Mbps | Middle Avg 107.75 Mbps | Max 112.43 Mbps | 6% Variance

TiP Data Points:: 78.2 Mbps, 110.13 Mbps, 110.54 Mbps, 110.22 Mbps, 112.35 Mbps, 110.94 Mbps, 109.73 Mbps, 108.86 Mbps, 104.93 Mbps, 107.85 Mbps, 112.43 Mbps, 108.24 Mbps, 105.44 Mbps, 106.26 Mbps, 105.66 Mbps, 103.85 Mbps, 102.3 Mbps, 102.02 Mbps, 47.84 Mbps

More Stats:: https://testmy.net/quickstats/CA3LE https://testmy.net/compID/69431494760

Test Time:: 2014-02-14 11:22:36 Local Time

Location:: Monument, CO US >> Destination:: Dallas, TX US

1MB Download in 0.08 Seconds - 1GB Download in ~1 Minutes - 1816X faster than 56K

This test of exactly 204800 kB took 16.502 seconds to complete

Running at 455% of hosts average (Comcast Cable https://testmy.net/hoststats/comcast_cable)

User Agent:: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/32.0.1700.77 Safari/537.36 [!]

The default testing method, which has been the core test for TMN since at least 2001... is still the default because I'm not here to make you feel good. I'm not here to overstate your speed or try to give you the best number possible. I'm here to give you the truth and hopefully help you improve performance. The linear test is much more helpful in getting the most from your connection. Tune your connection to perform better on that test and everything will perform better ... the other guys also get paid by the ISPs, whose interest do you think they have at heart? The CEO of speedtest.net is the former CEO of speakeasy.. an ISP (ref wikipedia). I wonder who his buddies might be... lol. TestMy.net is a third party, I'm not affiliated with ISPs. ..... take a minute and think about how much money a large ISP can save by making their users feel like their connection is running better. Then tell me that a third party opinion isn't necessary.

Besides, TMN is more than an internet speed test, it helps you discover other issues that other speed tests consistently don't pick up on. It truly reflects your computers performance not just internet performance (if your connection is fast enough). It's helped me position my router to perform the absolute best across my entire home, discover that modems were bad, realize that MTU and RWIN needed to be adjusted, it's even helped me realize that I was working with slow or degraded hard drives... drive swap and TMN results boosted. Because it's my business to test my competition I always compare, WAY more often than not these issues are not noticed, real physical issues that degrade performance... unnoticed. And when you're testing on a server that's down the street (even if it's 300 miles away) not taking into account normal peering and routing, how realistic is that? TMN tests you under real world conditions. Like I said, I'm not here to sugar coat your speed. I just hold up a mirror.

... there is so much more but I've already written more than most will read. If I've got you thinking... take the time to browse around TestMy.net, read what I've written... read what others have written too. Understand how it works then if you want to put it under the microscope, I invite you. You'll quickly realize that I've done my homework and I'm really here only to try and help consumers get the most. I've developed what has become TestMy.net for 18 years, if you like it please take 2 seconds to share it with a friend.

let us know what OS you're running and your network configuration and we'll try to help you improve your score.

Link to comment
Share on other sites

Thanks for the reply.  I thought shaperprobe was linear, but i actually couldnt tell as it didnt show any data being transferred under the TCP connections in resource manager.

 

I actually did read all that stuff, and im a programmer, just not super familiar with networking.  Im just trying to figure out why all of a sudden within the last 2 weeks I can barely stream.  Comcast keeps saying everything is ok, ive tried 3 different modems, router or no router, wireless and wired, different computers and i cant figure anything out.  Sometimes everything is good, sometimes i dc every 30 seconds. I thought for sure it was a wireless issue, but when i dragged out all the cat5 cables and put them all around my house and hooked up my gaming pc nothing changed. i was able to stream on wireless without a problem for months, with half the upload speed that i have now.

 

Anyways, I changed the server to the alt cdn one and now im getting results that I can believe to be reasonably accurate.  Like 60mps down out of 100.  I can believe that since I think there is a problem with my connection (i know there is a problem with my connection actually just not what) and the linear makes it slower. Plus i pointed speedtest.net to like other servers in dallas and got numbers not too differing.  That I really dont understand.  Why would my numbers drop off so much going from a server right by me to a server in dallas?  Bandwidth isnt a function of distance afaik.... What does that indicate?

 

It was just kinda shocking at first to be like running tests and see literally 10mpbs here and 120 on speedtest.net (connected directly to modem, router is only 100mpbs).  That was with my old computer though. So I apologize for that.

 

I just want to know wtf is wrong with my connection so I can tell my ISP, because they just tell me everything is fine.  It shouldnt be MTU or whatever because everything was working with all the same stuff in my house and then just stopped working fine out of the blue, but i downloaded TCPoptimizer or whatever it is and ran that.


So it seems like at my absolute max that ive gotten here I get about 60% up/down of what im supposed to be getting, and what my isps tool would report.  What can cause that?  Im at my wits end here its going on like 3 weeks of me trying to troubleshoot this. thanks.

Link to comment
Share on other sites

I'm not exactly sure what network configuration means.

 

Im running windows 7

right now i have a modem wired to a router with 25ft cat5 cable and then another 10 feet coming out of the router to the pc. 

 

These problems have persisted through 2 routers and 3 modems, and just started happening out of no where sometime after january 26th (the last log i could find where i was able to stream at 3mpbs for like 8 hours straight, now I get disconnected all the time or drop frames at 2.5mpbs but its somewhat intermittent).

 

If you want other info heres screens from tcp optomizer, i just changed these settings today though to what the program recommended:

 

http://i.imgur.com/GQjvs1D.png

 

thanks

Link to comment
Share on other sites

Those drop offs are obviously what's affecting your streaming.  You're getting decent speeds, they just aren't consistent.

 

First thing I would do is check the cable and all of the fittings, trace back all the way to where the cable company's line comes in.  Kinks in the cable, bad splitters, bad barrel connectors can all cause intermittent signal loss.  A kink or sharp turn in a coaxial cable messes with the impedance of the cable. ... I've seen many times also where people have stapled the cable to baseboards (to tidy it up) and the staples are so tight it pinches the cable... often overlooked... or where the cable has been fished under the carpet and ONE baseboard tack has managed to prick the cable.  A little nail acts like a miniature antenna, feeding air signal into the cable.  I've seen things like that cause transmission to cut in and out like that.

post-2-0-33321000-1392416077_thumb.png

post-2-0-24367100-1392416078_thumb.png

post-2-0-19057800-1392416079_thumb.png

Link to comment
Share on other sites

From the modem:

 

post-169191-0-11012800-1392424105_thumb.

 

When the cable company came after i started having problems they found 1 line or whatever that was giving off noise and it was because it was an old coupler. They replaced it and that went away, but my problem didn't get any better.

 

The problem happens on more than 1 computer although i dont try to stream on the other one, its a barebones pc running xp.  It happens both wired and wireless.  The cable company said i was getting really good signal in the room with the coax to the modem and they couldnt find anything aside from what i mentioned at where the cable comes into the house.  I believe the cable is split once when it comes in and that is where the internet gets split off. then the other side of the split gets split some more but its after the internet is already split.

 

I mean would it really be in my house just all of a sudden after months of it working just fine?  I was on the phone with the cable company and the guy said he was pinging me with 128k packets for like 10 minutes straight and only a tiny amount got lost, well within acceptable range.  When the connection drops like that shouldnt that lead to packet loss?  Cuz i was dropping half my frames while i had the stream running while i was on the phone with him.

 

This is definitely new and it doesnt just affect my stream its just the easiest place to see theres a problem, it obviously has an effect on these tests and i was playing a game earlier today (havent played a game that needed a constant online connection since the problem started) and was definitely feeling the troughs just being in the menus having them lock up for a little bit.  This never happened before and i played like 400 hours of that damn game.  It all just started sometime after january 26th.  Thanks

Link to comment
Share on other sites

If I might interject a thought - many times after heavy freeze / thaw cycles, moisture seeps in one of a plethora of small spaces. Freezes and expands in the process. Add a bit of wind then warmer tempuratures with load, and any node will fail. Transferring data streams to secondary routes. Which in turn overloads already heavily used networks. Which is not the point of my post.

 

The switching to and fro the failover networks due to an anomoly of the same sort will cause precicely what you are describing.

 

Just a few thoughts.

Link to comment
Share on other sites

This is an interesting thread as I am new to this testmy.net site and have been using speedtest.net for a long time trying to keep my isp TWC honest in what they are suppose to provide.  Not trying to hijack the thread but should I post my similar questions here?  I am probably many steps behind UnfriendlyBG but would like to use this site to improve my internet speed to what TWC said I purchased?  If there is a newbie thread on how to run the tests and perform PC turning please point me to it.  thx

 
Link to comment
Share on other sites

If I might interject a thought - many times after heavy freeze / thaw cycles, moisture seeps in one of a plethora of small spaces. Freezes and expands in the process. Add a bit of wind then warmer tempuratures with load, and any node will fail. Transferring data streams to secondary routes. Which in turn overloads already heavily used networks. Which is not the point of my post.

 

The switching to and fro the failover networks due to an anomoly of the same sort will cause precicely what you are describing.

 

Just a few thoughts.

I can attest to this experience 6 months ago as I was running speedtest (have not discovered testmy.net) and the speed dropped from Down 20M to about 5M.  TWC did a lot of remote testing and wasted a lot of time having me reboot the cable modem and finally by pass the router by direct connect CB -> PC.  Then they agree to send out a tech but threaten me if the problem is after their cable demark then I have to pay for the service call and fixing the problem.  I told them to come and get it as I am fairly certain it is not my house ethernet cabling.  Long story short, I shadowed the tech when he came out and after checking signals inside he found degraded signal and traced it to the junction box in the sidewalk under ground across the street from my house sprayed by a sprinkler daily. The junction panel was wet from the outside and the rubber seals were cracked.  So he replaced the junction box, changed all the cable seals and then retuned the signal from the street junction box about 1500 feet away.  So I got my 20M down back but this will happen again some years from now when the replaced sidewalk junction box / connector seals will give out again.  I guess it is job security for the tech as he has no control over the people across the street running their sprinkler and neither do I.     

Link to comment
Share on other sites

well that's very interesting to me. I've had it in the back of my mind that it almost corresponded with some snow we got, again 1/26 is the furthest i can go back, im not sure the exact day.  It snowed around those days i just looked it up and its also barely been over freezing, hitting 33 degrees a few times.

 

Also when I called comcast originally for this problem to set up a tech to come out, they called me back that night and said they were doing some work in the area because people were complaining about low speeds or something and asked if the problem was fixed.  It actually did work that day iirc until like 3am or something then it started up again.

 

What exactly is a node? is that something by my house?  According to my streaming logs, theres a slight correlation with time of day and the intermittent connection but it can happen at any time.

 

Why would the guy pinging my computer with 128k packets for 10 minutes not see a problem when im seeing a problem at the very same time with my frames being dropped?

 

Basically I need to know EXACTLY what to tell comcast because they really dont want to work with you. All they will do is send the cable guy out to my house to just check the signal strength and all that maybe replace my modem (for the 3rd time).  Those guys arent like networking gurus. The last guy wanted to just plug 1 router into the other router and i was like im pretty sure that doesnt just work like that.  I was like well how would forwarding ports work in that situation? And he didnt know what forwarding a port was, so obviously they just are dudes that set up your connection and stuff, theyre not going to solve this for me most likely.

 

It took me like 6 calls to even get connected to a tech department.  I had like 2 people just outright tell me I cant talk to any technical people on the phone which is completely untrue it turns out.

 

The guy I did speak to in what i believe to be the 'home networking department' did sound like he knew what he was talking about though.  He was the guy pinging my computer.  We did have a dyscrepancy over something though. I was saying how downloading was affecting my uploading. Like id start a speed test and the frames on my stream would drop. And he said theyre completely seperate that shouldnt be a problem. And I was like well if youve ever used p2p filesharing you know that if your upload is too high you can make your download crawl. and then he said well thats because it uses 1 port and that this situation is different and some technical gargon so i just differed to him on that because he sounded like he knew what he was talking about. But again, couldnt 'see' anything wrong with my connection he only really seemed concerned when I told him i was getting like 30mpbs downloads but i overslept and missed the appointment for the 2nd tech to come out.

 

edit: And if the problem is away from my house, why does my connection bottom out when uploading? Wouldnt windows show my traffic as going outbound at a constant rate and then the packets get lost? I dont really know how all this works, and where packetloss would come into play or not.  But i just ran a speedtest and got 2.2 mbs up with like half the time it not uploading any data. Im supposed to get 20mbps!

Link to comment
Share on other sites

  • CA3LE locked this topic
Guest
This topic is now closed to further replies.
×
×
  • Create New...