Search the Community

Update: The Multithread Speed Test doesn't just open more threads, it can also test servers simultaneously and return a single result. Get a single number for a world-wide speed test. Select one or all of TestMy.net's servers. Hey testers! I came up with something new yesterday. It came together better than I had imagined. It's actually a twist on an old technique I used to use. So, the normal download speed test streams the data, as you would read a document. This new method loads elements that are able to be simultaneously downloaded, in a way that allows your browser to max out. I achieve speed far greater using this method, confirmed by readings from the controller. It taxes my connection so hard it causes my Pandora to pause and skip. That's when you know you're pulling all it's got (confirms that I'm pulling the same speeds I get from Usenet and torrents... still 50 Mbps below what Cox advertises for my package). The speed builds faster and is held longer... but I've also seen in my test bed where the opposite is true, the configuration and the browsers optimization and performance come into play for sure. If for example the browser is only set to allow a limited number of concurrent connections, then performance on this test will obviously suffer. This in no way is meant to replace the download test I have now, they are two totally separate methods and can both be used in their own way to find issues. They're both correct... they just request the information in different ways so the result is different. The normal test finds things that multithreading would hide. The information from the multithreaded test can be just as useful. In my opinion comparing the results of the two test types can lead users to discover and resolve more issues easier. Okay, I blabbed for long enough. For anyone that stuck around... Turn on multithreading - switch the server or follow that link again to turn it off. This only works with the download speed test and disables TiP when enabled. Realize that this method could be more resource intensive. That's the point to a test like this... you want to tax your resources. It crashed one of my VPS servers early in its development... although I haven't seen a problem since. Having said that, I also saw zero issues when testing on desktop, tablets and smartphones. Test with this option and let me know if you find anything weird. -D