|
|
|
Having upgrade to BT 900, like others I'm not getting the full 900Mb. BT of course just say it is above the minimum speed.To be fair they did fiddle about behind the scenes and fixed the upload being stuck at 50Mb
As recommended I have set up a direct PPoE between my PC and the ONT, using a cat 6 cable, as a test to eliminate the smart hub 2.
The problem I am having is that the speed-test is slower this way than when connected through the hub.
Direct to ONT - 120 Mbs
SH2 - 600+Mbs
Upload is the same for both at 110Mbs
Tried with different cable and PC , results the same
This is how I set up the PPoE
Step 1 Press “Windows key +X”, and click on “control panel”;
Step 2 Then go to Network and Internet->View network status and tasks page.
Step 3 Click on Set up a new connection or network
Step 4 Select Connect to the Internet and click Next button.
Step 5 Click on Broadband (PPPoE).
Step 6 Insert username and password.
Any ideas why it is slower without the hub inline? I assume I have just done something wrong in setup but can't see it
|
|
|
Is your PC showing a gigabit network connection when connected to the ONT? Is the CPU maxed out during a speedtest?
Cerberus FTTP + pfSense + UniFi nanoHD
|
|
|
|
What are you using to speed test? Many test servers cannot cope with the full 900 Mbps. If you're seeing download speeds of 500-900 Mbps with upload of 100+ Mbps on the SH2 then nothing to be concerned about.
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
Any ideas why it is slower without the hub inline? I assume I have just done something wrong in setup but can't see it
What spec is your PC, and what security products do you have installed?
21 years of broadband connectivity since 1999 trial - Live BQM
|
|
|
|
It could be Windows network drivers, anti-virus software, browser limitations, many things.
I use a Macbook Pro as my main machine. If I run a browser-based test speedtest to speedtest.net then I can see in Activity Monitor that the CPU is maxed out at 100% while downloading; the measured speed bounces around and doesn't achieve more than about 270M.
However if I run the standalone speedtest.net app, then I get a flat ~305M which is what my (300/50) line is capable of, and the CPU isn't maxed out.
|
|
|
However if I run the standalone speedtest.net app, then I get a flat ~305M which is what my (300/50) line is capable of, and the CPU isn't maxed out.
Does that depend on the browser? Sounds to me as if your DMA is not working from Ethernet card to RAM.
Are you using WiFi or an ethernet dongle on the Mac? Which type of dongle, USB or Thunderbolt? Is your disk encrypted, which if the browser is writing to disk could add to CPU load.
I have a 4th gen i7 desktop running Win10, and gigabit connection to my router, and then gigabit to the Virgin Hub. The gigabit is an Realtek chip on my Asus motherboard. My disk is not encrypted, as its a large, heavy, tower chassis.
200 Mbps download, 20 Mbps upload. Almost no impact on CPU using a browser based speed test, peaked at 30% CPU for about 2seconds, but mostly 5% during the test on speedtest.net or thinkbroadband.com
I am using Firefox (version 81.0 currently).
21 years of broadband connectivity since 1999 trial - Live BQM
Edited by jchamier (Thu 24-Sep-20 21:15:53)
|
|
|
DMA doesn't work using a browser but works using the app?
The CPU utilisation is the browser. I see the same thing.
Building better networks, not just faster ones.
|
|
|
True DMA can’t be the reason.
Are you also using a Mac ?
21 years of broadband connectivity since 1999 trial - Live BQM
|
|
|
|
Hi,
Yes it shows gigabit connection and no it isn't maxed out
|
|
|
|
A variety including Fast.com which BT tell you to use
|
|
|
|
I7 with 16Gb Ram. I tried it with the AV disabled and it made no difference. When other people do this they either get the same or faster speeds not significantly less which is what I do not understand.
|
|
|
Fast.com tends to over read, based on looking at network card throughput and figures displayed.
This is BT900 over wireless https://www.thinkbroadband.com/speedtest/16009896835...
Would switch on the PC but its late, an old one is http://tbb.st/1598039865460825655
Posting the link to the results from the Thinkbroadband test can sometimes give some clues to issues.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
|
Connection is via Thunderbolt to a Belkin TB2 dock, and gigabit wired ethernet from there. Using iperf3 to a local server I can max out the gigabit easily.
The browser I use is Chrome. Firefox is slightly worse - it gives 230-250M.
I suspect the bottleneck is because of the external 4K monitor I use, connected via the same thunderbolt dock to DisplayPort. During the test, the top CPU processes are "Google Chrome Helper" and "Google Chrome Renderer".
SSD is encrypted, but a speedtest website isn't going to be writing all its downloads to disk (or shouldn't be)
Anyway, I was just making the point that the bottleneck is often the client machine, especially if using a browser-based speed test.
|
|
|
Anyway, I was just making the point that the bottleneck is often the client machine, especially if using a browser-based speed test. It is an interesting finding, given MrSaffron has shown faster than gigabit working on the TBB tester in a browser. There does appear to be quite dramatic performance differences for browsers on different Operating Systems, which is very interesting.
21 years of broadband connectivity since 1999 trial - Live BQM
Edited by jchamier (Fri 25-Sep-20 09:44:44)
|
|
|
https://www.thinkbroadband.com/speedtest/16010244607... using Firefox
MacBook Pro via Ethernet using Anker powerexpander+ 5
Nothing special and was literally first speed test I've ran today
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
Nothing special and was literally first speed test I've ran today Thanks.
21 years of broadband connectivity since 1999 trial - Live BQM
|
|
|
Try running the thinkbroadband test and posting the results. this will tell people more to judge what is going on. If you post both, Mr Saffron will be able to see more of what the potential difference is.
Edited by kitcat (Fri 25-Sep-20 14:30:25)
|
|
|
Yes indeed I am using a Mac.
Building better networks, not just faster ones.
|
|
|
Yes indeed I am using a Mac.
I just got my service the other day, very happy indeed but I was wondering also on the speed difference & using a 16'' mac with Belkin usb-c Ethernet.
https://www.thinkbroadband.com/_assets/speedtest/but...
|
|
|
|
Ok, I will post them tomorrow, one though the Hub and one connected to the ONT.
As my original post, I am not trying to see why I cannot get 900 - that will come later- but why the same speedtests are slower with a direct connect than through the hub when the perceived wisdom is that this is the way to go to see whether the hub is the problem or the fibre/ont.
|
|
|
why the same speedtests are slower with a direct connect than through the hub when the perceived wisdom is that this is the way to go to see whether the hub is the problem or the fibre/ont.
The fact that you're already getting 600+ Mbps on the BT router suggests there's nothing wrong with the service at BT's end. If you're not getting the same speed on a laptop/PC connected directly to the ONT then there is probably a limitation on the laptop somwhere, eg AV software, insufficient RAM/CPU etc. I wouldn't worry too much about this considering its normally your router which should be left connected to the ONT 24/7, not a PC. People usually test at the ONT with a PC if their router doesn't give them the expected speeds, not the other way around!
|
|
|
|
Thanks
|