Question about download speed?

[g-idk]

Hi, I was wondering which speed test I should be trusting?  The reason is I have internet 300 with the Hitron Modem and I have been consistently getting good download and upload speeds via speedtest.net by Ookla, anywhere from 270 up to 310 download and 16 upload.  When I use Shaw SpeedTest powered by Ookla however I consistently get 150-180 download and 16 upload, even when I run the test side by side.  So who is right?? Any thoughts or suggestions would be appreciated as I'm not a tech, and I find this very confusing.  

[mdk]

@g-idk -- if you access the Shaw Speed Test, you should see the text "powered by OOKLA" near top-centre, and something like:

Let it automatically choose the Shaw server closest to you.

Running the test will give you the speed of the "last NN kilometers" portion between that server and your cable-modem.

If you run two Speedtests simultaneously, using two computers, on your 300 Mbps service, then you would expect 150 Mbps into each computer. Correct?

[g-idk]

Hi @mdk thank you for responding.  I do let Shaw Speed Test automatically chose server, 99 percent of the time it shows Vancouver as the server of choice and I reside in Burnaby, so I presume that is the closet to me.  When I said running the test side by side I was referring to run speedtest.net in one tab and get results then open another tab and then run shaw speed test  and get those results, they are way way off by 100mb., by the way I don't have 2 computers.  So I've tried just going to Shaw speed test and getting results and then just going to Speedtest.net and getting results and like I said tried the 2 tabs open way.  Oh I should mention that it does the exact same thing with my Edge browser and Firefox same outcomes, Speedtest.net gives me 270-310 download and 16upload and Shaw Speed Test gives me a reading of 150-180 download and 16upload.  This is a head scratcher for me.  

[mdk]

@g-idk -- run speedtest.net in one tab and get results then open another tab and then run shaw speed test  and get those results, they are way way off by 100mb.

Compare to driving from Burnaby to a Starbucks in New Westminster to driving to a Starbucks in Whistler. Same product, but much less time to get to N.W.

Note that "speedtest.net" is several "hops" further away from you than Shaw's server in Vancouver:

   9 ms  rd1cv-be210-1.gv.shawcable.net [64.59.162.97]
   9 ms  rd1lp-be1.gv.shawcable.net [66.163.72.22]
  15 ms  rc1wt-be40.wa.shawcable.net [66.163.68.18]
  11 ms  six1.fastly.com [206.81.81.71]
  10 ms  151.101.194.219

So, when you run that test, you are at the mercy of the speed all those "hops", including some in WA(shington state). Correction: even though the Speedtest.net web-server is in the USA, the actual testing-server is the SAME one that the Shaw Speed Test uses. So, there should be no difference when running two consecutive tests. For you, does it make a difference if you run 5 tests consecutively: Shaw, SpeedTest, Shaw, SpeedTest, Shaw ?

> it does the exact same thing with my Edge browser and Firefox same outcomes,

This implies that there are no issues in either web-browser at running the "client" part of the software on your computer -- one web-browser is not "faster" than the other.

> Speedtest.net gives me 270-310 download and 16upload and Shaw Speed Test gives me a reading of 150-180 download and 16 upload.  This is a head scratcher for me.

Me, too. Experiment:

1. Launch the Windows "Task Manager"
2. Switch to its "Performance" tab
3. Click on the icon for your network connection
4. Minimize this window
5. Launch the Shaw Speed Test
6. Restore that Task Manager window
7. Observe the transfer-rate during the "download" section of the test
8. Use the Windows "Snipping Tool" to capture & save an image of the Task Manager window
9. After the test completes, minimize the Task Manager window
10. Launch the speedtest.net program
11. Restore that Task Manager window
12. Observe the transfer-rate
13. Again, capture an image of the Task Manager window
14. Compare the two images
15. Post the two images here, if they are significantly different

I hope this helps.

[g-idk]

Hi again @mdk , ok I think I did what you requested, I ran multiple test like you recommend but still the same outcome here is some snips for you.  this is speedtest.net 

this first one is SpeedTest.net

this one is Shaw Speed Test

This one is SpeedTest.net

and this one is Shaw SpeedTest.  I ran the test 6 times one after the other and get basically the same results each time.  I don't think my computer likes Shaw Speed test. If this is not what I was supposed to send you please let me know.  

[mdk]

@g-idk -- I have internet 300 with the Hitron Modem 

Those are good graphs. Thanks.

I have the same modem and the same "Internet 300" grandfathered plan, but I get slightly over 300 for download and slightly over 15 for upload.  So, your numbers for download should be around 300, not just over 200, and the graph should be a "plateau" with no "spikes" -- certainly the "upload" section of the graphs are plateaus, because Shaw is "throttling" your upload, to match your plan.

I think that you should contact Shaw, and ask them to remotely logon to your Hitron, to view the signal-strength, and other values, including "transmission errors", to see if you're getting a good signal incoming to the Hitron.

Can you try a different Ethernet cable between your computer and the Hitron?

Can you try a different coaxial-cable between the wall-outlet and the Hitron?

Do you have any other active wall-ports in your home (TV in a different room?) where you could temporarily connect the cable-modem, and rerun the tests.

Do you have a 1-to-2 coaxial-splitter, so that your PVR and your cable-modem each get "half" the signal coming-out of the wall-port?  Can you temporarily bypass that splitter -- by connecting the cable-modem directly to the wall-port?

Edit: I have a fast (3.2 Ghz, quad-core) desktop still running Windows 7. Here is the output from my Shaw Speed Test:

Definitely a "plateau", at over 320 Mbps. The second "block" is wider, because I changed the "update speed" to "fast" -- taking more measurements per second.

[g-idk]

ok @mdk , I will try your suggestions a little later today and get back to you, thank you for the help so far. 

[g-idk]

Hi again @mdk , ok I've done some troubleshooting as per your request;

1)  replaced ethernet cable (cat 5e), re ran test, same results

2)  replaced coxial cable between modem and splitter , re ran test, same results

3)  do not have any other wall ports in my apt. only the one in the living room 

4)  Yes I have a 3 way splitter since I have home phone, internet and tv.

5)  Yes removed splitter and plugged modem directly into wall outlet, re ran test, same results

6)  Installed new coxial cable from outlet, re ran test, same results

7)  I have an older dual core 2.8 ghz amd processor running windows 10

I think my computer doesn't like Shaw Speed Test  I just took a quick snip for you on what the websites are telling me after all the changing of cords I did. 

this above is SpeedTest.net, letting it pick server.

And this of course if Shaw Speed Test,  For the amount and what I use the internet for this doesn't actually affect me in any way, I have no problem streaming Disney, prime in full HD,  I just for the life of me cannot figure out how 2 different speed test that are supposed to be powered by the same software can be so out of whack, would my processor cause such huge differences?  Oh I haven't called Shaw Support yet I'll wait till after I hear back from you. 

[mdk]

@g-idk -- I am puzzled by the difference, and also by the mention of TELUS in the 1st graph. Any idea why it shows?

>  I have an older dual-core 2.8 Ghz AMD processor running Windows 10

Can you rerun the Shaw Speed Test, with the Windows Task Manager active, and showing the level of the processor activity?

My thought is that the Shaw Speed Test is single-threaded (not exploiting your 2nd core), but that Windows dynamically does 50% of the web-browser's processing on 1 core, and the other 50% of it on the 2nd core, resulting in a "50%" displayed value for the load on your processor.  If your processor is "too slow", then the app inside the web-browser is not running the app at its full speed, resulting in a lower displayed result.

It is curious that the previous graphs for the network usage were approximately the same, for either app. This implies that your processor is fast-enough to run either test at full speed -- limited only by the speed of your cable-modem.

[g-idk]

Hi @mdk , I did the processor graph for Shaw SpeedTest here is the snip

It stayed at 100% for most of it.  Same thing for SpeedTest.net it stayed at 100% for most of it as well.  About Telus showing up in the SpeedTest.net server it does that lots of times and sometimes it also says multi connections, ie. Telus plus 3 more.  Once in a Blue moon it will pick Shaw Server as best but very rarely.  Boy I'm at a loss to understand.