Multithread

[sirdir]

Hi!

 

I've read somewhere that testmy.net avoids using multiple threads to show 'the real speed' of a line instead of 'squeezing out the last bit'...

I don't really agree.

In my situation, testmy.net doesn't show 'my speed' at all. I currently own 4 Internet links and  load balance them with round robin. 

When I do a multithreaded download, I get up to 40 MBit/s, but as testmy.net doesn't multithread if I have bad luck my slowest line is tested and the result is only 2.6 MBit/s.

 

[CA3LE]

Hey man,

 

You've got an issue. Somewhere along the line something is limiting your single thread. You shouldn't have to multithread to max out your connection. I don't have to, other people don't have to.  Look at my signature.  My fastest speeds are from thousands of miles out, hundreds of Mbps. If I can do it... you should be able to.

 

It's funny that you say this right when I have a message in my PM box from sietec [host's sietec speed test in Miami FL]. He's pulling nearly 600 Mbps, over 70 MB/s, so fast I had to make an adjustment because the system thought his scores were an error.  I believe he's testing from Miami to Texas pulling those speeds. Again, if he can do it... why can't you?

 

He's also a host like you are... and he's really putting TMN under the microscope. Because he has the program on his servers he can scrutinize it big time. As can anyone. Part of my original intention of releasing this to other servers is that I wanted people to put it under the microscope, I stand behind what I've built. Many other people do as well.

 

Here's a quote from his last email.

 

As far as the logging problem, I kind of had a feeling that was what was causing it because you always have to make sure erroneous garbage is thrown out.  BUT, I never like to force MY opinion on other people's sites.  Besides, you don't build a site like TMN as a novice developer; that is for sure.  You have done an awesome job with this and I, being the forever skeptic, have tested on my end the validity of your measurements via an algorithm I wrote for measuring the instantaneous transfer speed in- and out- bound of a webserver and keeping a method of logging that result without any significant "averages" that always skew results for short lived transfers while at the same time not taking away from performance or building gigantic log files.  Your site has performed to errors which could only be attributed to my built in margin of error which results from short burst of very high instaneous xfer speeds over a few seconds.  In other words, I have found essentially zero discrepancy between what your results show and what I measure myself.  I cannot say that about any other speed test site, not one.  Hence the reason I have taken an interest in yours.  Awesome work.

 

"I have found essentially zero discrepancy between what your results show and what I measure myself.  I cannot say that about any other speed test site, not one."

 

That's a pretty heavy quote coming from someone posting speeds like this...

 

Once I make the upload test larger he'll be able to symmetrically pull 600 Mbps, I'm sure. Only lower because it's only 33MB.

 

This is the most accurate and compatible speed test online.

[tommie gorman]

Man, those are insane speeds. WOAH 

[CA3LE]

Hi!

 

I've read somewhere that testmy.net avoids using multiple threads to show 'the real speed' of a line instead of 'squeezing out the last bit'...

I don't really agree.

In my situation, testmy.net doesn't show 'my speed' at all. I currently own 4 Internet links and  load balance them with round robin. 

When I do a multithreaded download, I get up to 40 MBit/s, but as testmy.net doesn't multithread if I have bad luck my slowest line is tested and the result is only 2.6 MBit/s.

 

There is now an option to enable multithreading. This will be expanded to mirrors very soon. -- you still need to question why you aren't able to pull more speed without multiple threads... but this gives you a tool for testing with more than one thread.

[sietec]

Hi all,

 

I just wanted to put my two cents in here for what it's worth.  First of all, I have no interest in any online speed test.  I host mirrors for the two most popular speed test sites, one of which you are on right now.  My only interest is in helping people get accurate and informed information when it comes to measuring their internet speed.  I will probably upset some people with this post, but what I am saying comes from my 12+ years of experience in tweaking web servers and meticulously measuring the details of different transports.

 

For http (e.g. "internet" or web-browser compatible) speed testing, I can say with 100% certainty that you will not find a testing methodology more accurate than testmy.net.  The attention to detail and every little tidbit of information involved in http/tcp transport is second to none.  I can say that since I host a mirror on here and can measure the results from both ends.  For example, if you make a request on my box, apache keeps a log of the transfer information and precisely how long the socket was open and how much information (exactly) was transferred. The timing is accurate to less than 1/10000 of a second.  Therefore, I can measure the precise time it took to transfer, say 100MB from my box to your computer and determine from that the speed in which the transfer took place. 

 

TMN goes a bit further than most, in that you can see the actual speeds during the xfer as it fluctuates.  The other speed test engines take an average of your speed over the duration of xfer and throw away a significant amount of the data as "erroneous."  My question is, if it is erroneous yet contributes to the overall time it takes to transfer a file, it is NOT erroneous and should NOT be discarded as such.  This is where you will see significant differences on the "other" sites. 

 

Sometimes it is disappointing to see the real numbers your machine is capable of.  Believe me, I know.  I get upset when I think a measurement is wrong.  But, it gives me valuable insight into how to increase  the throughput and achieve better results.

 

Take a look at my max and averages

 

 

 

These come from my VNC (e.g. linux remote desktop) connections to one of my boxes using Firefox or other browsers to test the box.  The average has been skewed by testing all kinds of different settings (e.g. jumbo frames, different MTUs, browsers, servers, etc) and so could be much higher if I dedicated an account singly to testing with my best settings...but that would kind of defeat the purpose, in my opinion!  If you consider the overhead on the processor and network interface just to support my 1920x1080 VNC connection, then factor in that I am remoting into a server to test it's connection; furthermore that it is just a TCP link with limited threads that can be opened (due to the nature of the protocol), my box is capable of much more.  But when it comes to true INTERNET ability, I have never measured anything faster than I can through testmy.net.  Consider that the "other" sites use Flash and other visual enhancements, a limited amount of data and basically, nondiscriminatly discards data assumed to be erroneous, which site do you think will be more accurate?

 

I will soon be starting a blog regarding how to improve you internet connection speed and will post the link when I publish it.  I would like people to try before and after modifying system settings as well as the other speed test sites.  I say everyone should try the other sites as well and see where the discrepancy lies.  But if you are truly looking to improve your performance, in my opinion, you have found the spot (that is, unless you want to write your own code and try to start from scratch what has taken a decade to get where it is...good luck by the way).

 

Anyone with questions, please feel free to ask.  Understand, please, that I am often distracted and cannot immediately reply -- but I will reply.  Probably your best source of information is in the threads on this site, or directly from Damon.  He is very active on his site and generally answers anyones questions or posts fairly rapidly. 

 

Thanks for reading and forgive my spelling and/or rambling!  Good luck!

 

SIETEC

[CA3LE]

... 

 

I can say with 100% certainty that you will not find a testing methodology more accurate than testmy.net.

...

 

SIETEC

 

Wow man, good read.  At a loss of words, rare for me.  It's been a while since I felt this validated.  Sometimes it feels like an uphill battle educating people on why the site is not like any other.  I can type till my fingers bleed trying to explain it... but coming from me it doesn't mean as much.  I really look forward to reading that blog.

 

Thanks for taking the time to share your opinions.

[xs1]

lol sorry, but i'm not going to write a 3 page essay on this topic... although i could. But like "sirdir", the creator of this thread... with Verizon fios, i TOO have to multithread to get my maximum speeds. and no, i'm not talking about the results that it gives me at the end i  pretty colors and visuals. I do know how to accurately monitor & measure my speeds on a real-time basis. When i am downloading a 1.3 GB file , and i see that it ended at "9.62 Mbit/s" i know that is what 'im downloading at. Their is no amount of overhead or "burst" that is going to last over a gig's worth of download time. tesmy.net's system as well as various NDT servers give me terrible , awful results. - correction, well now that D implemented the multithreading i can NOW  max out my connection. But as the thread starter said, multithread is the only way i can achieve my max connection potential. Also proven in DC++. Maybe not so much a "problem", but the ISP's intentions for some ungodly-irritating reason?