CA3LE

There should be some sort of static URL we can use to download a large file to do a speed-test from a command line, such as Ubuntu server.
 
For example, we should be able to use wget and go to a static file on whatever test server we prefer.
 
(I.e.: dallas.testmy.net/static?size=5 which would download a 5MB file)
 
Thanks,
EBrown

The test should request the bandwidth that the user expected. This doesn't have to be the bandwidth they were actually (legally) promised, but what they understood they would get. This could be provided once for the Auto Test, and after each individual test. A default value could be stored in a user account, but the user would have to press a big button to confirm it each time.  This would provide three possible types of values:
User Entered/Confirmed: xMbps Profile default: xMbps Last Entered/Confirmed: xMbps I think this extra data is very useful.

Handbrake is awesome.  Especially for quick codec conversions.  Actually they do have a version for windows and even Ubuntu.  Download Handbrake -- You may also want to try VirtualDub.  Both are completely free.  
 
 
Plus, I love this... "Put down that cocktail, your Handbreak queue is done!"    

First, welcome back! Been a while since you've logged in. One of the first 1000 members, I'm glad to see you're still visiting.

Under certain circumstances the progress bar can lag... which doesn't effect the outcome of the test. I've seen this on VNC and on Android 2.x phones. When I've seen this it never pauses, then turns into a fluid progress bar (meaning that it moves smoothly). It will pause then JUMP... as if the browser couldn't keep up with the java used to update the progress. Regardless of how the progress bar updates the flow of data is uninterrupted. On the other hand you might have an issue that's causing this. If it pauses, moves, pauses... then the bar fluidly moves for the remainder of the test... I'd bet that there is something wrong.

First thing I would check would be MTU. You're on Bell South, which is DSL (right?)... PPPoE. Which means that the optimal MTU for your computer and router is 1492. Optimal MTU for cable Internet and most other providers is 1500. When I was on DSL for a while I was having a similar issue... my equipment was configured for my previous cable connection (I'm back on cable and don't miss fiber DSL at all, okay... the upload was pretty awesome. My connection on cable is much more stable and is especially much better for gaming. Plus, my nick is CA3LE... lol.) and changing the MTU resulted in a drastic difference. Like night and day. Check this out... CA3LE March to April 2010 -- you can see where I fixed it and stabilized my connection. (note that some of the results in that query are unrelated... TMN didn't have the filtering and identifying options back then so I can't narrow the search down to exactly the related tests. Only pay attention to the Qwest.net results). It's a very important setting and can have a profound impact on your performance. An easy way to tune your TCP stack in windows is with Speedguide.net's TCP Optimizer. But your router may also need to be configured correctly for PPPoE, it may have an MTU setting or may just ask you what type of connection you have. If you're coming from a different provider it's pretty common for this to be incorrectly set.

Here's a topic very similar to this one... Hanging while testing - I think his issue was resolved with a new NIC. So that could be the issue too.

A little side information, TestMy.net is different than the rest.
If you test on the majority of other Internet speed tests out there you may not see this effecting your results... because they've designed their tests to ignore the worst part of your results. I personally don't understand the mindset behind that, it shows where their interests lie... hint, they don't work for your best interest. TestMy.net on the other hand just calls it how it sees it. If there is a pause in the flow of information, it's reflected in the results.

Say that you download a 10MB file and it takes 10 seconds. Let's say that there is an extra 2 second pause at the beginning of that download. You might not notice it because once the download starts to flow it's going fine...... but does that mean that the 2 second pause never happened? Hell no! Even though it might have happened without you noticing, it still happened. In turn effectively dropping your speed for that download by about 17%. Now, let's say that you were downloading only 1 MB at the same speed, then the issue is compounded and effectively drops your speed by about 67%. I've had users, especially satellite users, that tell me that I should design the test to start the time after the speed has picked up... but wouldn't that defeat the purpose of the test? If your connection has pauses and blips, that's important information to know. It's the reason that you came here... right? You might stabilize at 8 Mbps (in our theoretical scenario) but that doesn't mean that's your real speed. Issues that cause that kind of problem make a HUGE impact on browsing speed, VoIP, gaming... etc. Ignoring that crucial information is exactly what most speed tests do, something TestMy.net will never do.

Admin:  First thank you so much or having this website.  I LIKE IT!!!  You do good job here.  I will patiently wait for some volunteers to have servers that are nearer to my home.
 
Again thanks for doing what you do.  I for one appreciate it as I am sure others do too...
 
Jim
===

I used to offer this option prior to 2005.  But I didn't like the way I did it before so I scrapped the idea for a while.  Today I'm releasing an option to those of you that would like to host your own speed test.
 
The first server to run this is Doolgo.com (My Son's website).  Hosted in New York.  Use the Doolgo.com Speed Test
 
This works with any web server running php 5 and Source Guardian.
 
If you'd like to become part of the beta download the file below and follow the readme, once you've uploaded the files send me a message with the website address you'd like to setup.  I'll put your information in the database and send you a key file that will get everything working.  
 
please request the latest build
Install in minutes... approved and running within 24 hours.
 
This is very early stage, more options etc are soon to come.  I will eventually build a GUI for the signup and management but for now just shoot me a message.  I'll update this thread as more information becomes available.

Starting today the east and west servers are gone. But opening up to outside servers should populate mirrors quickly. I've already got requests and I have a list of people who've requested that in the past to get to.

The new server in Texas is costing me more than three times the cost of all the mirrors. I had to cut some cost somewhere to justify the extra cost of the home server. I'm sure we'll already have a list going within a few days. That whole subset to the site will grow and give TestMy.net users many testing locations. There is also the auto routing option that routes via CDN.

When I went to TMN just now, in the Servers section up at the top (where you choose your testing location), the Eastern - and Western, too - servers are all gone.
 
Now the options are the following:
Central USA, Automatic Route, Europe, More Servers (?)
 
I click on More Servers to see if maybe you just moved those two I mentioned before in there, but all that's in there is just some public mrrors - the Eastern and Western servers have completely disappeared from the website.
 
Are they down, or did you take them off for whatever reason?
They were here yesterday, and I always use the Eastern server because it's routed closer to me (and then I sometimes test with the others to see how I compare).

wow...i like this...i was just looking on the wayback machine and saw where you used to do this

I run handbrake just fine on windows 7 x64 ultimate

Understood. I'm simply curious to see several tests from the central server which will certainly prove interesting results, which i do not think will reflect your latest / previous results.

Ok, here is what I did last night:
I did a bunch of tests in a row, on laptop, first wired directly to modem and then over WiFi.
25MB test download each time for both.

 
Wired               WiFi


Start
time         18:46               19:07


                                24.1MB/s             19.6MB/s

                                27.8                        3.4

                                27.6                        23.1

                                27.5                        22.8       

                                27.2                        3.2

                                3.4                          22.7       

                                3.4                          3.4

                                24.2                        3.3

                                25                           3.3

                                25.5                        22.3

                                23.6                        3.2

                             3.2                             22.7

                             3.3                             21.1

                             19.8                           3.0

                             18.3                           21.7

                             3.4                             22.9
 
End time             18:55                         19:18
 
sorry for format
 
So can anyone make nay sense of this???

<p><p>You might want to check your dd-wrt settings , you did mention you were running this correct ?</p><p>&nbsp;</p><p>There are a lot of settings in that guy, so many that to understand and properly utilize them all you might need to have a degree ( self taught or not ) !</p><p>&nbsp;</p><p>I might get to the basics and chuck the router for a minute, reboot the modem and connect the machine directly. And only one machine for this test.</p><p>&nbsp;</p><p>Run a traceroute to testmy.net, ping testmy.net ~</p><p>To ping a host...
Windows Start > Run > type CMD [enter] > from the command prompt type ping google.com and/or ping testmy.net Linux / Mac Under Applications/Utilities open 'Terminal' > type ping google.com and/or ping testmy.net Paste those results here.</p><p>To run a traceroute...
Windows Start > Run > type CMD [enter] > from the command prompt type tracert google.com and/or tracert testmy.net Linux / Mac Under Applications/Utilities open 'Terminal' > type traceroute google.com and/or traceroute testmy.net Paste those results here.</p><p>&nbsp;</p><p>Then run a few tests, but use the central server only for this test.</p><p>&nbsp;</p><p>First simply run any test that is at least the cap of what you pay for, run this a couple times to see the consistency. Run the upload test as well from the central server , at least the cap of your upload, and run this a couple times to note the consitancy.</p><p>&nbsp;</p><p>Please post the result, either copy/ paste the bbcode or html given after the test. As well as the ping and traceroute.</p><br /><br /><p>Hello forum bug --- ca3le please check html perms on special bbcode </p></p>

What speed should your connection be ?

I'm building a new desktop computer. I wasn't sure about how PCI Express 3.0 works. Right now the video cards im going to use are 2 Radeon HD 6850's. But in the future I would like to get a 7850 or better. I noticed the 7000 series cards say PCI Express 3.0. Does that mean that they require a motherboard with PCI Express 3.0, or they just take advantage of 3.0 if you have it but work with 2.0? I'm doing a AMD CPU build and I can't even find a motherboard that says it supports 3.0. Was just wondering about this. Thanks guys.

That's what I was wondering. It seems kind of weird AMD wouldn't have a chipset that could support it. My guess is that its backwards compatible with 2.0 it just doesn't move data as fast as 3.0. But yeah it would be nice to know for sure.
 
Edit: Just saw TriRan's post. Thank you for the info man. That is what I was thinking. Glad to know though. And yeah I would think that the difference from 2.0 to 3.0 wont really make a difference for a while.

3.0 video cards will work fine in a 2.0/2.1 motherboard all that will happen is the cards bandwidth will be limited by the motherboard as a bottleneck if you will I doubt you would even notice the difference though

I just a couple weeks ago went looking into this.  Does AMD even support true 3.0 yet?
I know a couple TMN members have been neck deep in this recently and should reply soon.

I was guessing eletrical interference. 

It's possible that the variation you're seeing is real and normal, although if you're seeing all the way from 4 Mbps to 20 Mbps, that seems a bit much. Try selecting the test file size - go to the Download Test menu and pick say 25 MB for a test file size. Run the test 3-5 times and see if you get more consistent results. You can try it using the test server near you, and the automatic server selection-at the very top of the page.

And pick say 1.5 MB for the Upload test size when you run that one.

Report back with some results

If you still have big variations, you may have an equipment problem inside or outside your house.

Good idea goodwin
 
I've seen the same thing with wireless, sometimes there are other reasons aside limited bandwidth for this, such as 900mhz cordless phones, or wireless camera's, things of this nature. At the same time , the path to the wireless connection may be getting interference from one of the above, there could be limitations or QOS settings in the wireless configuration, that are throttling certain packets or ports. Such as P2P, streaming media, VoIP and gaming port ranges.

First, welcome to TestMy.net and you're welcome!  I'm glad you like what I've built.  I've been at it for a very long time, building this site is my passion.
 
If you're using NoScript or RequestPolicy or any other Internet security software you have to whitelist tmnstatic.com.  I serve my static content through a different domain name for a couple of reasons.  First, being under a different domain name it tricks your browser into opening more threads to load the page.  This makes the page load faster.  Second, it's served under a 'cookieless domain' so I cut down on request headers which saves a little bandwidth because the cookie information isn't added to the headers of all the requests on that domain name.  If I didn't do that unnecessary cookie data would be added to the request headers for every request.  So, you load the page and all of the images (for instance) would have extra unnecessary information added to every single request.  This is a relatively small savings but extrapolated across so many images and other requests the savings starts to add up.  Then if you extrapolate that over ALL of TMN's traffic... then it really starts to add up.  I try to save as much bandwidth and requests as I can, it's easier on the client and easier on the server.  I want my bandwidth to go to testing connections not loading pages.  Things load much faster with the optimizations I have in place.   ... unrelated but I also minify and gzip and cache whenever I can.  Saving even more bandwidth and speeding up page loads.  I also serve tmnstatic.com from a CDN, so when something is requested from that domain name you download the information from one of many datacenters that may be closer to you than TMN's home server.  This speeds up loading time even further.
 
If tmnstatic.com is being blocked for any reason... resources needed for the pages to load properly won't be loaded.  So the page output will be totally off.      ... obviously, tmnstatic.com is a trusted domain, as long as you trust TestMy.net.
 
As long as you don't have a third party program blocking that and your browser is set to default settings you shouldn't have any problems.  Every browser I've tested, under default configuration without add-on's like yours, has loaded perfectly for me.  Even very old, outdated browsers.

First, if you don't plan on adding value to conversation... leave.  Simple posts saying "hi" that contain no true value to the forum may be seen as spam and removed.  Tell us something worth typing... share your speed, share your experience or tell me what you think of what I've built.  As long as you're a real person posting real content you have nothing to worry about.  Bots on the other hand don't last long here.

We'd love to have you introduce yourself as a new member, so please click here to make a post to the New Members board that you're currently in.  That way we can all give you a warm welcome and answer any questions you may have.  But please, if you have a question about your system, connection or any computer related question please make those posts in the board that suits the topic best.  If you don't know where your topic should go just post it to General Discussion and one of us will point the topic in the right direction for ya.

To see a list of the boards you should make your posts to all you need to do is go to the Forum Default Page.  Make sure you take note of the sub-forums that are with-in some of the forums.  These sub-forums can sometimes give you more specific help.

HAVE FUN DAMN IT!
Most importantly... enjoy yourself here, this community is full of not only helpful self proclaimed (and proud of it) computer nerds it's also full of people having a great time!  Many people become addicted to our forums so I'd just like to warn you... you may be with us till the day you die, but at least you'll make a ton of really kick ass friends along the way.

- Damon

phree c00ki3s... that's cool.



Don't you just love how I tell you the truth here... it helps rather than hurts.

You've just seen first hand the difference between TMN and other Interent speed tests. This is for multiple reasons.
The other guys test your speed under the most ideal circumstances... they've removed many of the variables that are the Internet. Personally I believe that you're not here to know your speed under ideal conditions... you're here to test your Internet speed. How can you test your Internet speed if you're not even truly going out on the Internet? The tests that providers and the people that work for the providers best interests host are hosted within your providers network... TMN is a test that stands for the best interests of the consumer.

For more details please read https://testmy.net/legit-speed-test.php

Other speed tests run through third party protocols. These often bypass issues that you may have with your browser and OS configuration. They're also, in their own admission, often guessing... and adjusting the results based on guesses because of the poor protocol choice they've made. You didn't come to test your Internet for a guess, right?

For more details please read https://testmy.net/ip...la-speed-tests/


TestMy.net is the only speed test that works the way it does. Your results along with countless others prove my point... results elsewhere often can't be trusted. Especially if the provider is involved, which is often the case. They have a vested interest in showing you that your connection is fine and that the problem is with the website you're trying to visit. It cuts down on service calls and saves them money. Personally, I could care less about how many service calls those guys have to send out... your connection should work the way that your provider advertises. I'm here to hopefully keep them honest. TMN is a third party with no vested interests in the outcome of the results.

Having said that, this wasn't Cox's fault. This was a problem with settings on your end... a fault of your operating system. As you found, it's a simple fix with SpeedGuide's TCP Optimizer. Seems to me that the providers would keep much happier customers if they just sent them to TMN in the first place rather than having their customers have to figure it out on their own. It used to be standard operating procedure before many of them realized that hosting their own tests made them look better. Although I still think testing out to the Internet is better information the information on how your connection runs within their network is also useful... but only if the test being used is based on sound testing principals. If they're using Flash... like all of them are... the principals aren't sound. You saw the difference first hand, help me spread the word... people need to know this stuff.

Yeah, they have more time on their hands and tend to care more about the customer.  Subcontractors want to hurry up so they can get to the next job to make more money.
 
Here's a good example.  I went out on a job in North Scottsdale AZ... there had been 6 previous calls to this mansion before I got the call.  All the previous 'techs' were subcontractors.  I could see work that had been done to the drop and the side of the house... but inside I notice that there is a very old connector behind the wall plate... and the cable is RG-59, with copper shielding.  Already off to a bad start.  I go up into the attic (huge by the way, I could stand up with about 6 ft of headroom... straight up mansion) and see a litteral rats nest of cables.  Chewed all to shit by mice.  Some were so bad there was 10 feet of cable with nothing but the core left.  I thought to myself, "how the hell did they not see this!" it was instantly obvious as soon as you popped your head in there  ... because they weren't doing their job, that's how.  I called my dispatcher and had her give all my other jobs to my buddies because I had an all day job on my hands.  I ended up doing a nearly 30 outlet rewire on this place.  Luckily the attic was big and easy to work with.  The lady was cool too, she hooked up a big blower fan at the entrance to blow A/C in there for me.  So the job went really smooth.
 
... we charged her nothing for the entire job.  These days, they'd charge for every single outlet.  But obviously, if you have 30 outlets... money isn't a problem for you.     -- Funny thing is, if the contractors had seen and repaired that, they would have made BANK.  Probably a couple grade for that job.  But they were too lazy to get into it.  You should have seen the D-Mark.  I had a two way splitter, split to two 6 way amps, split to a bunch of 3 and 4 way splitters.  By the time I left she had one of the best pictures I'd ever seen on analog.
 
Contractors don't care about the customer in my experience.  They care about how many easy jobs they can knock out real fast.  I'd hear them talking to each other at the office as I got my equipment from the warehouse every day, "Dude, I did 20 data installs yesterday!" --- sorry but if you knock them out THAT fast, something isn't being done right.    ... and here comes the in house tech to clean up the mess.  Having said that however, Cox does follow up on them and holds them accountable for mistakes.  Some of them lose their contracts, but just because they lose their contract doesn't mean that there isn't a bunch of houses out there with mis-wired cable that they did.  It's not like firing them fixes all the issues they caused.
 
Also, prewired coax done by homebuilders is crap.  Brand new houses I'd go to that were getting service for the first time... first thing I'd do is cut off ALL the fittings that they did and throw them and the crap splitters they used in the trash.  Electricians know electrical, not cable.  I really just wished that they would just run the lines and leave the fittings up to me, hell it would save themselves money too.  They almost always did it wrong and it was just easier to do it all over and not risk a callback.  Probably one of the reasons my callback rate was one of the lowest in the state.